EARLY DETECTION OF LUNG CANCER USING DEEP LEARNING

Abstract

Lung cancer remains one of the leading causes of mortality worldwide, primarily due to late-stage diagnosis and the complexity involved in detecting early pathological changes. This project presents an automated system for the early detection and classification of lung cancer using deep learning techniques applied to histopathological images. The proposed system utilizes a Convolutional Neural Network (CNN) based on the VGG-19 architecture to classify lung tissue images into three categories: benign, adenocarcinoma, and squamous cell carcinoma. Unlike traditional machine learning approaches such as Support Vector Machines and Random Forests, which rely on handcrafted features and often yield limited accuracy, the CNN model automatically extracts hierarchical features from images, enabling more precise classification. The methodology includes data collection, preprocessing (such as normalization and augmentation), model training, and performance evaluation using metrics like accuracy and confusion matrix analysis. The results demonstrate that the VGG-19 model achieves superior performance in detecting and classifying lung cancer types, with improved accuracy and reduced misclassification rates. Additionally, a user-friendly interface is developed to allow medical professionals to upload images and obtain real-time predictions, thereby assisting in faster and more reliable diagnosis. Overall, the system highlights the potential of deep learning in enhancing early lung cancer detection and supporting clinical decision-making.